DIY Economiser for OEM Contactors

[Alibro]

Hi guys
Like many others here I have an OEM battery pack that came with contactors so I plan to save some cash by reusing them.
The issue with using OEM contactors is they do not come with economisers therefore draw more current from the 12V system than is ideal.
When the car is turned on the 12V system will be at 14V and the OEM contactors will draw over 700mA each. This may not be an issue for the battery but could result in the contactors burning out so is something I want to avoid.
I originally considered using a dc buck/boost module to lower the voltage to 10V but I have concerns that using such a low voltage to switch the contactors may be as bad as having them running at 14V.

Anyway I have been chatting to another member on his project thread about this and we may have a simple solution so I started this thread to discuss it further.
Below are the messages from his project thread.

Any thoughts on my idea of using a buck converter to reduce the voltage down to 10V for the contactors. I'm using the VW ones in my build and didn't like the 700mA they were drawing at 14.4V. I read somewhere that anything over approx 500mA would shorten their life somewhat.

The simplest solution I can think of would be a nano and a uln2003 wired in parallel (or bigger fet, preferably through hole fet) that would give 100 percent for first second then ramp down quickly afterwards OBS you could go for lower powered chip than a nano but they are cheap on Ali and don't need to faff with programmer.

That's an interesting idea.
It occurred to me that using a lower voltage like 10V to trigger the contactor cause problems long term? They operate fine down to 6V or 7V but I figured 10V would be safer but I like the sound of your idea better.
Any chance of a circuit diagram?

the diode would be worth a try first...but here is a simple diagram of my quick and dirty one that could be done on prototyping board.

pwm.png

If your interested in a doing a DIY pwm contactor driver I am quite confident i could do something with an Attiny85 and some through hole components.

Forgive my ignorance, I'd never heard of a ULN2003 before and my electronics knowledge is extremely limited but I can see they are very useful.
Would they be OK with being controlled by a PWM signal to reduce the voltage?
Would we need to put a cap across the output to smooth things a bit?
I guess we could use one to control two contactors by using half the input/outputs for each?

They should be ok to control voltage, here a quick google video where someone uses one as a motor speed controller.


My personal preference is to use all of the channels of uln2003 in parallel as they are relatively cheap (the chinease smd versions can be brought for $0.11 from lcsc although a real through hole one can be brought for £0.41 from mouser), i have read that their real world useable maximum is 100mA/Chanel, so using all would probably be a good idea.

or just go full lunatic and use a big transistor like BDX34C still only £0.61 (https://www.mouser.co.uk/ProductDetail/ ... OetQ%3D%3D)

The capacitor shouldn't be needed i would think as the inductor itself will act a bit like a capacitor and the uln2003 has a flyback diode built in.

I would also use an analog input pin (with protection) and a voltage divider to measure the 12v line to the pwm would be adjusted for when the dcdc is on/off or low battery etc.

here is a fully featured schematic with protection etc.. which i stole off previous work...

pwm.png

if you want me to check these and make a quick pcb design let me know

[Bigpie]

The OEM contactors from the VW GTE's I've been using without economisers, they don't seem to need them, though do get warm.
I've got some gigavacs that do need them and I've been using a 10W resistor with capacitor for a couple of years viewtopic.php?f=17&t=531

[Alibro]

The OEM contactors from the VW GTE's I've been using without economisers, they don't seem to need them, though do get warm.
I've got some gigavacs that do need them and I've been using a 10W resistor with capacitor for a couple of years viewtopic.php?f=17&t=531

LOL, before posting I did a search but didn't see this.
Any thoughts on which option is better?

[Bigpie]

No idea, this way is simple and doesn't have any switching noise but does generate heat and waste It was initially just a temporary thing until I came up with something else, but not thought about it since.

[Alibro]

Just in case someone reading this is slightly more clueless than me this is my understanding of how the two options work.

Capacitor and resistor
A capacitor is connected in parallel with a resistor and the two are in series with the contactor coils.
At the moment you turn on the 14V to the contactor the capacitor is a short circuit across the resistor and the coils receive full 14V
When the cap charges it becomes open circuit and the coil is now in series with the resistor so some voltage is dropped across the resistor lowering the voltage at the coil.
Simple but wasteful and the resistor could overheat and burn out.


ULN2003 and an Arduino Nano or similar.
The Arduino is programmed to generate a PWM voltage starting high then ramping down. This is fed to a ULN2003 (stepper motor controller) which has full 14V input. The output from it will initially be full 14V but will ramp down with the PWM feed from the Arduino.
More complicated so more could go wrong but is also more controllable.

It's been 40 years since I did electronics at college and I was rubbish then so if this is wrong please correct me.

[arber333]

Some time ago i found those economizer drivers. Might be worth to look upon.
At least driver was purposeful made to drive DC contactors.
viewtopic.php?p=39989#p39989

[Bigpie]

That actually looks like a nice solution. Might throw something together on EasyEDA later. EDIT found this https://www.buildcircuit.com/pwm-soleno ... ng-drv103/

[arber333]

That actually looks like a nice solution. Might throw something together on EasyEDA later. EDIT found this https://www.buildcircuit.com/pwm-soleno ... ng-drv103/

Nice! Just send gerbers for mfg...

[Alibro]

That actually looks like a nice solution. Might throw something together on EasyEDA later. EDIT found this https://www.buildcircuit.com/pwm-soleno ... ng-drv103/

They are selling the same thing here for around 35 Euro plus shipping
https://www.electronics-lab.com/project ... ng-drv103/
I think I'll have a go at doing it myself.

[Zieg]

I have done some trials with the drv series and haven't had much success. Damn things were sometimes opening or closing the contractors without an actual change in signal. Even the TI tech support couldn't figure it out. They also told me their datasheet was wrong and the input signal could be higher than 5v.

I've also been working on an RC economiser which seems to be working well. Everything was designed in KiCad, happy to share the files later in case anyone wants to have a look and see if they can figure out what the hell.

[snelly]

Here is some info,

https://www.ti.com/lit/wp/slvaf35/slvaf ... %253D17245



I used a Logo8 which are about £100 new to manage contactors' on a passat gte pack with panasonic connector's if any one the code please message me

https://na.industrial.panasonic.com/pro ... ies/121461

[Pete9008]

I have done some trials with the drv series and haven't had much success. Damn things were sometimes opening or closing the contractors without an actual change in signal. Even the TI tech support couldn't figure it out. They also told me their datasheet was wrong and the input signal could be higher than 5v.

I've also been working on an RC economiser which seems to be working well. Everything was designed in KiCad, happy to share the files later in case anyone wants to have a look and see if they can figure out what the hell.

20230524_072221.jpg

Nice board, be good to get it working.

I'd be a little surprised if there is a problem with the chip, TI/Burr Brown are normally pretty good. I'm wondering if your decoupling cap is a little too far away from the pins and supply transients are an issue. Might be worth bodging a cap directly across the chip's Vs and ground pins?

(to be fair the data sheet does say you can apply higher voltages to the input as long as they are current limited to no more than 2mA)

[Zieg]

Gave it a try but no dice. I have never done actual pcb design before though, so perhaps there are still hidden errors..

Here's the project file if anyone is interested.

drv103 pwm economizer.zip

(55.76 KiB) Downloaded 46 times

[Pete9008]

I think the transistor input circuit is your problem. You have a pnp input transistor referenced to the +12V rail but it looks like the collector and emitter are backwards (TBH I can't see how it works at all like that??). It's usually best to try and use the transistor symbol on the schematic rather than a box as it makes it much easier to read.

What control voltage are you driving it with?

Apart from that it all looks pretty good - not at all bad for a first PCB design

Edit - If you do revise it then it would be worth moving the ICs down the board so that their output is right next to flyback diode which is positioned right next to the connector pins. It would reduce the length of the high current path and simplify the tracking a bit.

Edit2 - Something like this:

I've changed to a potential divider on the input instead of the transistor, changed the connector pinout to ease the routing and put GND on a plane (probably ought to add one on the front too to keep the construction balanced or track ground instead). The 12V net would need to track across the top of the board or be another plane.

Edit3 - oops, just realised I deleted the power pad via, that will need to go back on!

[MattsAwesomeStuff]

They are selling the same thing here for around 35 Euro plus shipping

Just FYI, you can buy an entire contactor with a built in economizer, out of an EV, for $30 ($15 on sale) at Batteryhookup.

900vdc, 500A.

https://batteryhookup.com/products/te-c ... a-0-900vdc

[Zieg]

Well shoot that's embarrassing! I'll double check the actual part I used - thought I had probed that part of the board and are the time believed it was working. The intent was to convert the negative signal coming from the OpenInverter to a positive signal as needed by the DRV103. Is there a way to do the same thing with a potential divider?

The voltage the board sees is just the 14v from the dc/dc converter (or, I suppose, the 12v battery before the contractors close).

Thank you for taking the time to look at it! Sorry about that transistor symbol - it's just what came in when I added the specific component I planned to use, and I didn't know how to change it.

[Pete9008]

Be worth double checking the transistor, I could well be wrong!

Can the OI output handle the contactor current? If so you could just wire the chip as permanently enable and wire the supply to the OI output. It will then give a delay on power up (think there is a diagram showing it in the data sheet).

Otherwise you will need a transistor, it's just a little tricky to make sure it behaves in all permutations of supplies enabled and outputs on off. Will this board share the OI 12V supply or could it be powered while the inverter is unpowered?

[Zieg]

Hmm, I believe it could handle the economised contactor current, but I'm not sure if it will like the initial spike before the pwm kicks in. It will share the power supply, yes. My hope was that it would be able to work with a variety of contactors and boards so as to benefit as many people as possible.

I'll probe the transistor again and see what it's doing. I may also solder in a small relay and a pulldown resistor just to see if that works. Feels a bit wrong, but if it works I might have to go with it.

[Pete9008]

In that case I'd spin the transistor so that the emitter is to the 12v rail, add a 1kish pull up across the input to 12v to make sure it is off when the inverter output is off (not sure what the leakage current on the OI drivers is). You will also need a resistor between the collector and the drv chip input to make sure the current is limited to 2ma, guess something around 20k.

It's also possible that the transistor as is has managed to switch enough current to damage the chip input (transistors do work backwards, just pretty badly), if so a new chip might help.

[Alibro]

Just FYI, you can buy an entire contactor with a built in economizer, out of an EV, for $30 ($15 on sale) at Batteryhookup.

900vdc, 500A.

https://batteryhookup.com/products/te-c ... a-0-900vdc

Good point,
The only thing with buying used contactors is they may not be good. I bought two from a seller in Israel a couple of years ago and when I went to use them one had a high resistance that seemed to vary randomly so I was not very confident in the other.
I have two of these https://s.click.aliexpress.com/e/_DDx6p9b in my current build so we will see how they work out. They looked identical in every respect to the genuine used ones I ended up chucking and funny thing look just like the ones Batteryhookup are selling except made in China instead of Mexico.

[Pete9008]

This is the input circuit I'd use to invert the input sense. The 2k2 is there to make sure leakage current can't turn it on and the additional 4k7 to reduce the voltage and limit the current into the drv103:

Edit - might also be worth adding another not normally fitted resistor that bypasses the transistor. That way it could be built without the transistor circuit fitted so that it also accepted a non inverted 12v input or even a TTL level signal.

[Zieg]

Okay, awesome, thank you! I'll work on those changes asap and see what testing I can do with the boards I already have. Then I'll post another revision and see what everyone thinks before ordering more parts.

[Zieg]

Edit- Oop, just realized I slightly moved one of the mounting holes by accident, will fix before I order anything...


Rev 1.2 is good to go, I.. think? I reworked the transistor and added a spot to solder a resistor (SIGINV) that would bypass the transistor for people using a + signal. Good thinking!

Sorry the schematic is still a big damn mess, I'm still learning the software and moving stuff around is not yet intuitive to me.

As for the pcb layout, it was a complete rework with the following points:
-I got rid of the mounting hole I had between the chips (idea at the time was to use it to hold a heatsink, but I don't think it's actually necessary).
-Smaller screw terminal
-Bypass cap positioned right on the DRV103's doorstep
-Ground plane!

If nobody objects, I plan to order the PCBs later this week. Only $11 for three at Oshpark!

[Pete9008]

Were you able to test the changes on your last boards? Did they fix the problem?

PCBs looking better but still not quite there yet. The transistor connections don't quite look right.

Do you have the kicad libraries installed right? On mine when I go to add component I can choose that exact transistor and it pulls in the right schematic part and footprint automatically. It wouldn't surprise me if the drv103 is there too. I find the right click menu is very useful, for parts it contains things like rotate and horizontal and vertical flip which can help make the diagram more readable.

Regarding the heatsink what current would you like it to work up to? Either way its worth adding a small plane shape connected to the central device pad to help draw heat out of it.

Re. The holes and connectors it's worth locking them (right click menu or properties window) one you have placed them, makes it a lot harder to accidentally move them while doing the rest of the board.

Are you fixed on the connector pinout? Swapping a few connections would make the board routing much simpler.

Happy to help with, or make, the changes if that helps?

[Zieg]

Hm, I thought I was following what you showed. Maybe I got it wrong again..

Here's the exact part I was using. I was trying my best to use components that I could easily find the footprints for. https://www.digikey.ca/en/products/deta ... -F/1934466